Expression of beta 1B integrin isoform in CHO cells results in a dominant negative effect on cell adhesion and motility.

(1)

29 July 2021

AperTO - Archivio Istituzionale Open Access dell'Università di Torino

Original Citation:

Expression of beta 1B integrin isoform in CHO cells results in a dominant negative effect on cell adhesion and motility.

Terms of use:

Open Access

(Article begins on next page)

Anyone can freely access the full text of works made available as "Open Access". Works made available under a Creative Commons license can be used according to the terms and conditions of said license. Use of all other works requires consent of the right holder (author or publisher) if not exempted from copyright protection by the applicable law.

Availability:

This is the author's manuscript

(2)

Expression of /1B Integrin Isoform in CHO Cells Results in

A Dominant Negative Effect on Cell Adhesion and Motility

FioreUa Balzac, Saverio E R e t t a , A d r i a n a Albini,¢ A n t o n e l l a Melchiorri,¢ Victor E. Koteliansky,§ M a s s i m o G e u n a , * L o r e n z o Silengo, a n d G u i d o Tarone

Department of Genetics, Biology and Medical Chemistry, * Department of Biomedical Sciences and Oncology, University of Torino; Istituto Nazionale per la Ricerca sul Cancro, Genova, Italy; and § Laboratory of Physiopathology of Development, Centre National de la Recherche Scientifique 1337 and Ecole Normale Superieure, Paris, France

Abstract. The integrin subunit/~IB, a B1 isoform with a unique sequence at the cytoplasmic domain, forms heterodimers with integrin t~ chains and binds fibronectin, but it does not localize to focal adhesion sites (Balzac, E , A. Belkin, V. Koteliansky, Y. Balabanow, E Altruda, L. Silengo, and G. Tarone. 1993. J. Cell Biol. 121:171-178). Here we analyze the functional properties of human BIB by expressing it in hamster CHO cells. When stimulated by specific antibodies, ~IB does not trigger tyrosine phosphorylation of a 125-kD cytosolic protein, an intracellular signaling pathway that is activated both by the endogenous

hamster or the transfected human/$1A. Moreover, expression of BIB results in reduced spreading on fibronectin and laminin, but not on vitronectin. Ex- pression of BIB also results in severe reduction of cell motility in the Boyden chamber assay. Reduced cell spreading and motility could not be accounted for by preferential association of BIB with a given integrin ot subunit. These data, together with our previous results, indicate that BIB interferes with B1A function when expressed in CHO cells resulting in a dominant negative effect on cell adhesion and migration.

C

V.LL adhesion and motility requires the coordinated interaction of the fibrous network of extracellular matrix proteins and the intracellular cytoskeleton (Wang et al., 1993) bridged by plasma membrane receptors of the integrin family (Hynes, 1992). Interfering with any of these three elements results in disruption of adhesive cell-matrix interactions. This can be achieved by digestion of extracellular matrix proteins with proteolytic enzymes, by disassembling the actin skeleton with specific drugs or by blocking integrins with specific antibodies. The transmembrane linkage between the extracellular matrix proteins and the cytoskeleton occurs at focal sites on the plasma membrane known as focal adhesions (Burridge et al., 1988; Gei- ger and Ginsberg, 1991). Immunofluorescence analysis shows that extracellular matrix proteins, integrins, and cytoskeletal proteins, specifically colocalize at focal adhesions (Burridge et al., 1988; Geiger and Ginsberg, 1991; Luna and Hitt, 1992). The molecular interactions between these components, however, is not fully understood. Inte- grins are heterodimers of/~ and o~ subunits consisting of a large extracellular domain, a transmembrane domain, and a short cytoplasmic tail (Hynes, 1992). The extracellular domain of the integrin heterodimers bind to matrix proteins in Address all correspondence to Dr. Guido "Ikrone, Dipartimento di Genetica Biologia e Chimica Medica, Universit~ di Torino, Via Santena 5bis, 10126 Torino, Italy. Telephone: 39-11-670-6679; FAX: 3941-663-4788.

a cation-dependent manner and in many cases recognize the tripeptide sequence Arg-Gly-Asp (Ruoslahti, 1991). The cytoplasmic tails interact with cytoskeletal elements. Ex- perimental evidence for the interaction of the cytoplasmic domain orB1 integrin with talin and with o~-actinin has been reported (Horwitz et al., 1986; Otey et al., 1990). Other molecules may be important in mediating integrin-actin interaction, these include tensin, paxillin, and zyxin (Turner and Burridge, 1991). Moreover, molecules with enzymatic or regulatory functions are localized in focal adhesions and may have a role in integrin-mediated signal transduction; these proteins include specific isoforms of protein kinase C Oaken et al., 1989), pp60 ~ (Rohrschneider, 1980), as well as p125FAK (Schaller et al., 1992). The latter is a cytosolic tyrosine kinase that is specifically activated when integrins are occupied by their ligands or are bound by specific antibodies (Burridge et al., 1992; Guan and Shalloway, 1992; Kornberg et al., 1992; Defilippi et al., 1994). The cytoplasmic domain of B1 plays a critical role in the association of integrin with the focal adhesion structure (Reszka et al., 1992) and in the activation of the p125FAK (Guan et al., 1991).

In this study we analyzed the functional properties of #IB, a cytoplasmic domain variant of B1A, where the last 21 COOH-terminal amino acids are replaced by a new sequence of 12 amino acids (Altruda et al., 1990). This form behaves similarly to ~flA in terms of fibronectin binding but has a re-

The Journal of Cell Biology, Volume 127, Number 2, October 1994 557-565 557

on October 12, 2005

www.jcb.org

(3)

stricted tissue distribution and does not localize at focal adhesion sites (Balzac et al., 1993). In this study, we show that B1B does not trigger the tyrosine phosphorylation of in- tracellular proteins, and expression of this molecule in Chi- nese hamster ovary (CHO) t cells results in a dominant negative action on cell adhesion and motility.

Materials and Methods

Constructs and Transfections

Stable transfectants of CHO cells expressing the human integrin/~IA or ~IB were obtained as described previously (Balzac et ai., 1993). Briefly, a 3.5- kb EcoRI fragment of the/3113 containing the entire coding sequence was inserted into the EcoRI-cloning site of the SV40-based expression vector pECE (Ellis et ai., 1986). The full-length eDNA for the human ~IA cloned in the pECE vector was a kind girl of Filippo Giancotti (Giancotti and Ruos- lahti, 1990). CHO cells were cotransfected with 20/tg of the plasmid con- raining/~1 eDNA and 2/~g of pSV2-neo (Southern and Berg, 1982), and neomycin resistant clones were selected in HAM's I712 medium with 10% FCS and 800 #g/ml of G418 (G1BCO BRL, Gaithersburg, MD).

Flow Cytometry Analysis

Transfected cells were detached from culture plates by incubation in 5 mM EDTA in PBS (10 mM phosphate buffer pH 7.3, 150 mM sodium chloride) and washed twice at 4°C in PBS with 0.1 mM EDTA and 1 mg/ml of BSA. The cells were then incubated for 1 h at 4°C in the same buffer with saturat- ing concentrations of the monocional antibody LM534 to the human BI integrin (see below). After washing, the cells were incubated 45 min with fluorescein-labeled affinity-purified secondary antibodies (Sigma Chem. Co., St. Louis, MO), and analyzed on the flow cytometer Facs-Star Becton Dickinson (Oxford, UK), equipped with 5 W argon laser at 488 nm. Five thousand cells per sample were analyzed.

Antibodies and Immunoprecipitation of lntegrins

The following mAbs reacting with extraceUular epitopes common to the two isoforms of human/31 were used: mAb TS2/16 (a gift from Sanchez-Madrid, Hospital de la Princesa, Madrid, Spain); mAb GI2 (a gift from Luciano Zardi, Istituto Scientifico Tumori, Genova, Italy) and mAb LM534 (a gift from Filippo Giancotti, Department of Pathology, New York University, NY). None of the above antibodies reacts with hamster/31, mAb 7E2 reacting with hamster integrin/31, but not with the human molecule, was a generous gift of Rudy lniiano (University of North Carolina, Chapel Hill, NC). Antibodies to cytoplasmic sequences of c~ 3, u5, and t~V integrin subunits were prepared in our laboratory and previously characterized (Defilippi et at., 1991).

lntegrins were immunoprecipitated from cells labeled with 125I. Label- ing of membrane proteins with 125I was performed as described previously (Rossino et at., 1990). Briefly, cells were released from culture dishes by 5 mM EDTA treatment in PBS and washed three times by centrifugation. Ceils were suspended in PBS containing CaC12 (1 mM) and MgCI2 (1 raM) and labeled with 1 mCi of 12sI in the presence of lactoperoxidase (200 #g/ml) and H202 (0.002%). For immunoprecipitation, labeled cells were extracted for 20 min at 4°C with 0.5% Triton X-100 (BDH Chemicals, Poole, Dorset, UK) in 20 mM Tris-HC1, pH 7.4, 150 mM NaCI (TBS) with 1 mM CaCIz, 1 mM MgCI2, 10/~g/ml leupeptin, 4 txg/ml pepstatin, and 0.1 TIU/ml aprotinin (all from Sigma Chem. Co.). After centrifugation at 12,000 g for 10 min, extracts were incubated with the specific antibodies for 1 h at 4°C with gentle agitation. Soluble immunocomplexes were bound to protein A-Sepharose beads (Pbarmacia, Uppsala, Sweden) and recovered by centrifugation. After washing, bound material was eluted by boiling beads in 1% SDS (Pierce, Rockford, IL) and analyzed by (6%) SDS-PAGE in the absence of reducing agents. The radioactive proteins were visualized by fluorography with sodium salicilate (Chamberlain, 1979). The amount of antibody necessary to quantitatively precipitate all labeled molecules present in the cell extract was determined in preliminary experiments. The endogenous hamster B1 was immunoprecipitated with mAb 7E2 while mAb LM534 was used for the transfected human ~1. The t2sI radioactivity pres-

1. Abbreviation used in this paper: CHO, chinese hamster ovary.

ent in the immunoprecipitated/31 was determined by cutting the corresponding electrophoretic band and counting in a gamma counter.

Western Blotting

For Western blotting integrin complexes were immunnprecipitated from un- labeled cell extracts with u-specific antibodies. Immunoprecipitated material was separated by SDS-PAGE and transferred to nitrocellulose using a semi-dry apparatus (Novablot, Pharmacia, Uppsala, Sweden) according to the manufacturer's instructions. The blots were incubated I h at 42°C in 5 % BSA, washed with TBS (150 mM NaCI, 20 mM Tris-Cl, pH 7.4), and incubated overnight in 10 pg/ml B1 mAbs in TBS, 1% BSA. mAb GI2 that specifically reacts in Western blotting with human BI and n-tAb 7E2 reacting with hamster/31 were used. The blots were washed three times with TBS, incubated 2 h with peroxidase-conjngated anti-mouse Ig (Sigma Chem. Co.) and washed two times. Bound antibodies were visualized by the chemi- luminescent detection method ECL (Amersham, UK). In the experiment shown in Fig. 7, the monoclonai antibodies were directly labeled with 125I by the Cloramine T method. The nitrocellulose filters were first incubated with mAb G12 to visualize the human/31. After stripping with 2% SDS at 42°C for 1 h to remove bound antibodies, the filter was reincubated with mAb 7E2 to visualize the hamster B1. Quantitation of the band intensity was performed by densitometry using a scanner connected to a personal computer (Biomed Instrs., Fullerton, CA).

Adhesion Experiments

Tissue culture microtiter plates were coated by overnight incubation at 4°C with the indicated concentration of purified matrix proteins and post-coated with BSA (Sigma Chem. Co.) for I h at 370C. Fibronectin was purified from human plasma by affinity chromatography on gelatin-Sepharose according to EngvaU and Ruoslahti (1977). Vitronectin was purified according to Yatohgo et al. (1988), mouse laminin from EHS tumor was obtained from GIBCO BRL (St. Louis, MO).

Cells at confluence were detached by EDTA treatment in PBS for 10 rain, washed twice in serum free DMEM, and plated on the coated tissue culture plates for the indicated times. To eliminate the contribution of protein synthesis and secretion, cells were pretreated 2 h with 20/~M cyclobeximide (Sigma Chem. Co.) to prevent protein synthesis, and plated in the presence of 20 #M cycloheximide and 1/zM moncnsin (Sigma Chem. Co.) to prevent secretion (Defllippi et al., 1992). Plates were rinsed twice with PBS to remove unbound cells, and adherent cells were fixed with paraformaldehyde and stained with Coomassie blue. Cell adhesion was evaluated by reading the absorbance at 540 am in a microtitre ELISA reader. The correspondence between the optical density and number of cells attached was confirmed in experiments showing that increasing the cell concentration in the original suspension resulted in proportional increase of the optical density.

To measure cell spreading, stained cells were photographed by choosing five random microscopic fields for each sample. The area of 150 cells per sample was determined using a computerized image analysis system (Bio- med Instrs.).

Boyden Chamber Migration Assay

Cell migration assay was carried outin Boyden chambers as previously described (Albini nt at., 1987). Briefly, CHO ceils and transfected clones were resuspended in serum free RPMI medium in the presence of 0.1% of BSA. Cells were then added (2 × 10 s ceUs/ml) to the upper compartment of the Boyden chamber and the lower compartment was filled with either RPMI medium containing 0.1% BSA as control, or with chemoattractants consisting of serum-free conditioned medium from 3"I"3 cells or of RPMI with 25 /zg/ml fibronectin. The chemoattractants were used at concentrations exert- ing maximal activity as evaluated in preliminary experiments. The two com- partments of the Boyden chamber were separated by a polycarbonate filter (8 #m pore size, Nuclcopore, Concorezzo, Italy) coated with gelatin (5 /Lg/ml, Sigma Chem. Co.). Cells were allowed to migrate 6 h at 370C in a humidified atmosphere containing 5 % CO2. Cells on the upper side of the filter were removed mechanically; cells on the lower surface of the filter were fixed in ethanol, stained with Toluidine blue, and 10 random fields were counted under a microscope at 160×. Each assay was carried out in triplicate and repeated at least three times. In preliminary experiments, the ability of the cells to adhere to the filters was evaluated by fixing and staining the upper side of the filters and observation under the microscope.

The Journal of Cell Biology, Volume 127, 1994 558

on October 12, 2005

www.jcb.org

(4)

Detection of Phosphotyrosine-containing Proteins

To analyze cell adhesion mediated by integrin monoclonal antibodies, tissue culture plates were first coated overnight at 4"C with 10 ~g/ml goat anti-mouse IgG (Sigma Chem. Co.), poslcoated with BSA for 1 h at 37"C and incubated with purified mAbs for 2 h at 370C. The optimal concentration for each mAb was established by measuring cell adhesion.

Cells at confluence were pretrcated 2 h with 20 t~M cycloheximide (Sigma Chem. Co.) to prevent protein synthesis, detached by EDTA treatment (5 raM) in PBS for 10 rain and washed twice in PBS containing 1 mM CaC12, 1 mM MgCI2. Cells were then resnspcnded in prewarmed DMEM medium, in the presence of 20/zM cycloheximide and 1 /~M monensin (Sigma Chem. Co.) to prevent synthesis and secretion of endogenous extracellular matrix proteins, and plated on tissue culture dishes coated with antibodies to either the endogenous (mAb 7F.2) or the transfccted (mAb TS2/16)/31 subunit. Cells plated on polylysine-coated tissue culture dishes were used as controls. Cells were spun at the bottom of the dishes by centrifugation at 1,000 RPM for 2 rain and incubated for 5 min at 37°C. The cells were washed twice with a stop solution (5 mM EDTA, 10 mM NaF, 10 mM Na4P207, 0.4 mM Na3VO4 in PBS), and detergent extracted in lysis buffer (1% NP-40, 150 mM NaCI, 50 mM Tris-Cl, pH 8, 5 mM EDTA, 10 mM NaF, 10 mM Na4P2OT, 0.4 mM NaaVO4, 10/zg/ml leupeptin, 4/zg/ml pepstatin and 0.1 U/rnl aprotinin) (all from Sigma Chem. Co.). Protein concentration was determined in each cell extract by the Bradford protein assay method (BioRad Labs., Hercules, CA, GmbH). Samples containing equal amounts of protein were subjected to 6 % polyacrylamide elec- trophoresis in the presence of SDS (SDS-PAGE) in reducing conditions. Proteins were transferred to nitrocellulose and processed for Western blotting as described above. The mAb P T ~ to phosphotyrosine (Sigma Chem. Co.) was used followed by a peroxidase-eonjngate anti-mouse IgG (Sigma Chem. Co.).

Results

Expression of [31A and #IB in CHO Cells Transfectants

CHO cells were transfected with the human/31B cDNA under the control of the early SV40 promoter and stable transfectants were isolated. Clones transfected with the human /31A isoform were also selected as control. CHO cells express the endogenous hamster/31A form but do not express /31B as assessed by immunoprecipitation with antibodies directed to the peptide sequence unique to ~ (Balzac et al.,

1993).

To determine the expression of/31, membrane proteins were labeled with lactoperoxidase catalyzed radio iodina- tion, and the transfected human and the endogenous hamster /31 proteins were immunoprecipitated from each clone by means of species specific monoclonal antibodies and analyzed by SDS-PAGE (Fig. 1). To quantitate the level of expression, the/31 bands were cut out of the gel and the amount of radioactivity was determined in a gamma counter. One control clone expressing human/31A (/~1A-22) and three independent clones expressing/31B (/3113-18,/31B-25, and/31B- 50) were selected and analyzed. As shown in Table I, clones /31B-18,/31B-25, and/31B-50 expressed increasing amounts of human/31B; the level ranges from 18 to 50% of the total/31 expressed at the cell surface (endogenous plus transfected) (Table I). The numbers in the clone name indicate the level of surface expression of the transfected protein (see below and Table I). Clone ~_B-50 was previously indicated as 18.2 (Balzac et al., 1993). In the control clone/31A-22, the transfected human/31A represents 22 % of the total. The ratio of transfected/endogenous/31 in/31A-22 and/31B-25 clones is comparable (Table I), moreover, the amounts of total/31 at the cell surface (endogenous plus transfected) in these clones are comparable to that present on untransfected CHO.

The relative levels of/31 expression in the different clones

Figure 1.

Immunoprecipitation of/31A and/31B in transfected CHO cells. Clones/31A-22 transfected with human/31A and clones/3113- 18,/31B-25, and/31B-50 transfected with human/3113 were surface radioiodinated and extracted with detergent. The endogenous (E) and transfected (T) /~1 were quantitatively immunoprecipitated, respectively, with mAb 7E2 to hamster/31 and mAb LM534 to human/31. The latter antibody recognizes an extracellular epitope common to/31A and/3lB. Lane i is material immunoprecipitated with non-immune (n. i.) mouse serum. After immunoprecipitation radioactive proteins were separated by SDS-PAGE and visualized by fluorography, a and/31 on the left indicate the positions of the two integrin subunits. A 100-kD band of unknown identity was also immunoprecipitated from the transfected clones with antibodies to the human/31.

were confirmed by flow cytometry analysis which also showed that all clones consist of homogeneous cell popula- tion (not shown).

~IB Integrin Does Not Trigger ~yrosine

Phosphorylation of a 125-kD Protein

It has recently been shown that interaction of/31 integrin complexes with their ligands or with a specific antibody lead to activation of tyrosine phosphorylation of intracellular proteins (Burridge et al., 1992; Guan and Shalloway, 1992; Kornberg et al., 1992; Defilippi et al., 1994). We thus tested whether transfected/31A and/31B were equally effective in this respect. To specifically trigger the response mediated by the transfected or by the endogenous/31 subunits, cells were plated on plastic dishes coated with monoclonal antibodies specific for either human or hamster /31. Tyrosine phosphorylation of cellular proteins was determined by Western blotting with phosphotyrosine antibodies. As shown in Fig. 2, in clone/~1A-22 both the endogenous and the transfected /31A integrins were equally capable to stimulate tyrosine phosphorylation of a 125-kD cytoplasmic protein. On the

Table L Surface Expression of 131A and 131B in

Transfected Clones

C l o n e s E n d o g e n o u s Transfected Total j31A-22 1670" 471 (22% of total) 2141 /31B-25 1570 542 (25 % of total) 2112 j31B-18 2233 498 0 8 % of total) 2731 /51B-50 822 825 (50% of total) 1647

* The data are derived from an experiment shown in Fig. I. The numbers indicate the counts per minute incorporated in the/31 bands. Values were obtained by cutting the corresponding bans on the SDS gel and counting in a gamma counter.

Balzac et al. Role of /JlB Integrin in Cell Adhesion 559

on October 12, 2005

www.jcb.org

(5)

Figure 2. Tyrosine phosphorylation of p125 protein in fllA and fliB transfectants. fllA-22 and fllB-25 cells were plated on dishes coated with mAb 7E2 to the endogenous l l (E) or with mAb TS2/16 to the transfected human fll (T). Controls were obtained by plating cells on polylysine- coated plates (PL). After detergent extraction equal amounts of proteins were separated by SDS-PAGE and transferred to nitrocellulose. Proteins containing phosphotyrosine were visualized by means of a specific monoclonal antibody (PT66) and chemiluminescence reaction. The region of the gel in the 96-160-kD range of molecular mass is shown.

other hand, in clone fliB-25, the transfected BIB did not stimulate significantly the tyrosine phosphorylation of the 125-kD protein (Figs. 2 and 3). The endogenous iliA, however, was active (Figs. 2 and 3), showing that the signaling pathway was not altered in this clone. The ability of the transfected human fllA to stimulate tyrosine phosphorylation in clone fllA-22 indicates that the human protein is active in the hamster cells and thus the lack of function of BIB is related to the structural difference between the two integrin isoforms. The same phenomenon was observed in clone fliB-50 (Fig. 3). Immunoprecipitation experiments showed that the 125-kD protein is specifically recognized by monoclonal antibodies to p125FAK (Sehadler et al., 1992), a cytosolic tyrosine kinase localized at focal adhesions (not shown). Expression o f fllB Results in Reduced Cell Spreading To test the adhesive response of the transfectants, cells were plated on dishes coated with antibodies specific for the transfected human protein. This allows it to directly compare

Figure 4. Adhesion of iliA and fliB transfectants to antibody-coated dishes. 8ia-22 (a) and #lB-50 (b) ceils were plated on dishes coated with mAb TS2/16 to the transfecled human fll and allowed to adhere for 1 h at 37°C. Cells were photographed after fixation and staining. In a, cells adhere to the substratum by means of the transfected iliA while in b cells use the transfected fllB. Bar,

100 #m.

Figure 3. Densitometric analysis of the 125-kD tyrosine-phos- phorylated proteins in fllB and ilia transfectants. Ceils were allowed to adhere to dishes coated with polylysine (PL) or with mAb 7E2 to endogenous fll (E) or mAb TS2/16 to transfected fll (T). The proteins containing phosphotyrosine of the clones fliB-50, fllB-25, and lia-22 were visualized by Western blotting (see Materials and Methods). Quantitation of the intensity of the band corresponding to the 125-kD phosphoprotein was assessed by densitometry.

adhesion mediated by either the BIB or by the i l i a isoform. As shown in Fig. 4, fllA-22 cells adhered and spread nor- madly, while fliB-50 cells failed to spread. Thus, when the fllB isoform is involved in cell-substratum adhesion, cells do not spread normally.

We then tested the adhesive properties of transfected clones on dishes coated with purified matrix proteins. Under these conditions both the endogenous and transfected fll par-

on October 12, 2005

www.jcb.org

(6)

Figure 5. Spreading of BIA and B1B transfectants on purified matrix proteins. BIA-22 (a, c, and e) and BIB-50 (b, d, and f) cells were plated on dishes coated with 2/zg/ml of purified human plasma fibronectin (a and b); 10/zg/ml of purified mouse laminin (c and d); and 10 #g/ml of purified human plasma vitronectin. After adhesion for 1 h at 37°C, cells were fixed, stained, and photographed. Bar, 100 #m.

Balzac et al. Role of S1B lntegrin in Cell Adhesion 561

on October 12, 2005

www.jcb.org

(7)

ticipate to adhesion. To quantify cell spreading, the area of _A the adherent cells was determined by a computerized image M analysis system. As shown in Fig. 5 and Table II, BIB- expressing cells spread less on fibronectin compared to ceils expressing fllA. The difference was even more pronounced on laminin (Fig. 5, c and d) and in both cases it was statisti- cally highly significant as determined with the Student's t test analysis. Increasing the fibronectin concentration in the coat- ing solution from 2 to 20 ttg/ml allowed more spreading of cells but did not abolish the difference between ~ B and B1A clones. Cell spreading on vitronectin, on the other hand, was comparable in fllA and BIB transfectants showing that expression of BIB does not lead to a generalized spreading defect.

When the number of cells adherent to dishes coated with fibronectin or laminin was measured, BIB transfectants were not significantly different from untransfected CHO or fllA- O 22 cells (Fig. 6 A). Only clone B1B-50 showed reduced adhesion on both substrates (Fig. 6, A and B). BIB-50 cells express high levels of BIB compared to the other two clones, suggesting that reduction in adhesion occurs only when/3IB expression exceeds a threshold level. The reduced adhesive properties of the/SqB-50 clone were also observed by measuring the kinetics of adhesion to a fixed amount of matrix proteins (not shown). On the other hand, all clones including BIB-50 cells adhered normally when plated on vitronectin- coated dishes (Fig. 6 C).

In conclusion, expression offllB leads to a specific reduction in spreading of CHO cells to fibronectin and laminin, but not to vitronectin. In cells expressing the higher level of

cell attachment was also affected. ~ _

BIB,

w

Expression of BlB Results in Reduced Cell Motility

To evaluate whether the expression of BIB in CHO cells had a consequence on cell motility, we evaluated cell migration through porous filter in a Boyden chamber assay. Cells were plated on top of gelatin-coated porous polycarbonate filters and allowed to migrate toward the bottom surface. Soluble fibronectin or medium conditioned by 3T3 cells were added in the bottom compartment as chemoattractants. As shown in Table III, B1B-25 and BIB-50 cells showed a significant reduction of migration compared to B1A-22 cells. Migration was inhibited by 75% and by 60% in clones BIB-50 and BIB-

Table II. Spreading of B1A and BIB Transfectants on

Extracellular Matrix Proteins

Clones Fibronectin Laminin Vitroncctin C H O 565 + 130" 285 + 105 570 + 75 B1A-22 555 + 128 276 + 112 580 + 90 fllB-25 480 + 106 148 + 29 574 + 62 p = 0.0084 p = 0.0001 p = 0.158 fllB-18 484 + 96 156 + 38 568 + 52 p = 0.0081 p = 0.0001 p = 0.140 fllB-50 477 + 85 150:1:54 555 + 70 p = 0.0088 p = 0.0001 p = 0.125

* Cells were allowed to adhere to dishes coated with 2 ttg/ml of fibronectin; 10 ttg/ml of moose laminin; 10 ttg/rrd of vitronectin for 1 h. To measure cell spreading, stained cells were photographed by choosing five random microscopic fields for each sample. The area of 150 cells per sample was determined using a computerized image analysis system. Values represent the mean cell area in arbitrary units +SD, The p value was calculated with a Student's t test.

'° [I -c.o

so 0 0 0,15 0,3 0,6 1,25 2,5 5 10 20 35 3O z s '~ 20 lO 5 40 fibronectin uglml

--o-clio

I

I , ¢ , ' f - I I i I I I | I 0,15 0,3 0,6 1,25 2,5 5 10 20 laminin ug/ml + CHO 35 + p tA-22 "~ 25 < 15 1o $ 0 i | | i t t i i 0 0,15 0,3 0,6 1,25 2,5 5 10 20 vitronectin ug/ml

Figure 6. Adhesion of/~IA and/~IB transfectants on purified matrix

proteins, Microtitre plates were coated with increasing concentrations of (A) human plasma fibronectin; (B) mouse laminin; and (C) human plasma vitronectin. Cells were suspended by EDTA treatment and plated in serum free medium for 1 h at 37°C. Cell adhesion was measured by reading the optical density after staining with Coomassie blue. <>, untransfected CHO; t3, fllA-22 cells; t,, BIB- 25 cells; o, fllB-50 cells.

25, respectively, compared to clone B1A-22. Also in this case, no differences in migration between B1A and BIB transfectants were observed on vitronectin.

Association of fllA and BIB with Hamster Integrin

Subunits

A possible explanation of the adhesive behavior of BIB- transfected cells is that B1B associates with specific integrin

on October 12, 2005

www.jcb.org

(8)

Table IlL Migration of/$1A and/$1B Transfectants in Boyden Chambers

Clones Chemoattractant

(lower chamber) /31A-22 BIB-50 /~1B-25

Conditioned medium 3T3, 122 + 3* 27 + 1 36 + 1

F N 25 #g/ml 92 4- 9 17 + 4 22 + 3

V N 50 #g/nil 50 4- 5 50 4- 3 ND

SFM! 8 4- 1 10 4- 2 12 4- 1

* Mean values + SD of number of cells (per microscopic field) migrated on the lower surface of the porous filter.

* Conditioned medium by 3T3 ceils.

} SFM, serum free medium.

Figure 7. Association of human and hamster 81 with hamster cx subunits. B1B-50 cells were immunoprecipitated with c~3, ¢x5, or c~V antibodies, and the heterodimers were separated by SDS-PAGE and transferred to nitrocellulose by Western blotting. The coprecipitated 81 was detected by incubating the filter with l~I- labeled antibodies specific for the human 81 (mAb G12). After stripping with 1% SDS, the filter was reincubated with 125I-labeled antibodies specific for the hamster/31 (mAb 7E2). Similar results were obtained with cells transfected with human BIA.

ot subunits generating a different ratio between integrin complexes normally present at the cell surface of CHO cells. By screening with a panel of c~ subunit specific antibodies, we determined that CHO cells express o~3/fll, cx5/fil, and ~xV; the latter one forms heterodimers with both ill,/$3, and/$5.

To determine the relative amount of/$1A and /$1B associated with each ~x subunit, we performed immunoprecipitation experiments from transfected cells using ¢x3, o~5, and oW antibodies; the immunoprecipitated material was run on SDS-PAGE and probed in Western blotting with antibodies to human B1 (mAb G12). The relative amount of the two B1 associated with each ol was then determined by densitometry. As shown in Table IV, the relative proportion of the human B1A and/$1B present in the three heterodimers was comparable, confirming that the two fll isoforms did not differ in their association with integrin ¢x chains. We also analyzed whether the human and the hamster fll molecules showed the same pattern of association with hamster ot subunits. The integrin complexes were immunoprecipitated with c~-specific antibodies and the associated fll subunits were detected by Western blotting with the antibodies specific for human (mAb G12) and for hamster/$1 molecule (mAb 7E2) as described in the Materials and Methods section. As shown in Fig. 7, the human and hamster/$1 subunits displayed similar association with a3 and c~5, while a slight preference of the human/$1 subunits for c~V was observed.

D i s c u s s i o n

The cytoplasmic domains of the integrin/$1 and ot subunits are highly conserved during evolution (for review see Sastry and Horwitz, 1993), while the primary sequence of the ex- traceUular domains is much less conserved. This indicates a strong evolutive pressure to maintain a given sequence in the cytoplasmic region for the various integrin subunits and

Table IV. Association of Transfected Human [31 Integrins with CHO Cell ct Subunits

Association of transfected B1 with:

Clones a3 ~5 ~V

BIA-22 32* 57 11

B1B-50 30 60 10

* The amount of transfeeted/31 associated with the endogenous ~x subunits was determined by densitometry after immunoprecipitation and SDS-PAGE and it is expressed as % of the total transfected protein.

suggests that the molecular interactions with cytoskeletal proteins and with proteins of the intracellular signaling ma- chinery pose rigorous structural constraints. Splice variants characterized by new cytoplasmic domain sequences have been identified for several integrin subunits including: /$1 (Altruda et al., 1990; Languino and Ruoslahti, 1992),/$3 (van Kuppevelt et al., 1989),/$4 (Tamura et al., 1990), or3 (Tamura et al., 1991), 06 (Hogervorst et al., 1991; Tamura et al., 1991), and or7 (Collo et al., 1993). On the basis of the considerations discussed above, these variant cytoplasmic sequences are expected to generate specific intracellular signaling and interact with unique cytoskeletal components. We have identified a/$1 variant,/$1B, characterized by a distinct cytoplasmic sequence (Altruda et al., 1990). Previ- ous data have shown that/$1B form heterodimers with integrin ot chains and binds fibronectin, but it does not localize to focal adhesion sites (Balzac et al., 1993). Here we further analyze the functional properties of this variant showing that /$1B does not trigger the tyrosine phosphorylation of intracellular proteins and has a dominant negative effect on cell spreading and motility.

/$1B is expressed in a tissue specific pattern and is always coexpressed with/$1A (Balzac et al., 1993). The coexpres- sion of the two isoforms hampers the functional analysis of /$1B in cells that normally express it. To circumvent this problem we transfected human/$1B in hamster ceils that express only the BIA isoform. In this system it is possible to probe the function of the two isoforms by using species specific antibodies directed to extracellular epitopes and it is, thus, relatively simple to assign functional properties to the transfected molecule.

We have used this approach to show that/$1B cytoplasmic variant does not trigger increased tyrosine phosphorylation of the cytoplasmic p125-kD protein. This molecule is a ma- jor component undergoing tyrosine phosphorylation in response to integrin-ligand interaction and it corresponds to the focal adhesion kinase p125FAK (Burridge et al., 1992; Guan and Shalloway, 1992; Kornberg et al., 1992). The lack of signaling through this pathway was not due to a gross defect of/$1B since this variant is exported at the cell surface and forms heterodimers capable of binding extracellular ligands (Balzac et al., 1993). Moreover, the endogenous B1A in the same transfected cells effectively triggers tyrosine phosphorylation, ruling out the possibility that the signaling pathway is nonfunctional in these transfectants. In addition, independent transfectants showed that human /$1A isoform is fully functional in signal transduction in the CHO back-

Balzac et al, Role of fllB lntegrin in Cell Adhesion 563

on October 12, 2005

www.jcb.org

(9)

ground. The lack of tyrosine phosphorylation of the p125 protein by/31B is, thus, due to the properties of its cytoplasmic sequence. These results are consistent with data reported by Guan et al. (1991) showing that specific deletion of the COOH-terminal region of the chick/31A, corresponding to the region spliced out in/31B, abrogated the ability to stimulate tyrosine phosphorylation of the 125-kD protein.

The /31B variant was also ineffective in promoting cell spreading when cell adhered to surfaces coated with antibodies specific for this molecule, This is consistent with our previous finding that BIB does not localize at focal adhesions (Balzac et al., 1993) and shows that cell spreading and signaling through tyrosine kinase require an intact 131 cytoplasmic domain.

Our data also show that expression of/31B in CHO cells has a dominant negative effect on cell spreading and motility. The inhibition of these adhesive functions is not due to un- desired mutation of the selected clones, since: (a) it is common to independent clones; and (b) it is restricted to fibronectin and laminin and does not occur on vitronectin. While fibronectin and laminin are recognized by integrins of the/31 class, vitronectin is preferentially recognized by/33 and t35 integrins (for review see Hynes, 1992). Thus, the effect of/31B is restricted to those substrata recognized by/31 class integrins. The/31B is functionally normal as far as the association with a subunits and binding to fibronectin are concerned (Balzac et al., 1993). We also show that/31A and /31B are indistinguishable for the association with or3, or5, and c~V subunits expressed in CHO (see Table IV). Preferential association of fllB with a given tx subunit would change the ligand specificity of the cells and lead to reduced adhesion to specific ligands. Moreover, comparative analysis of the human and hamster/31 proteins showed that the specie difference did not affect the association with hamster or3 and a5, although a slight preference of the human/31 for oN was detected. This, however, can not explain the distinct reduction of adhesion and migration on fibronectin and laminin. Thus, the dominant negative effect on cell adhesion is most likely explained by the fact that/31B competes with/31A for association with ot chains and binding to the extracellular matrix, but it fails to transduce mechanical or biochemical signals inside the cell.

It is interesting to note that significant reduction of cell spreading and migration is observed in cells where fllB represents less than 50 % of the total ~51 at the cell surface. Whether these effects can be explained solely by the competi- tion of fllB and/31A for heterodimer formation and binding to extracellular ligands is unclear. Other phenomena may contribute to the observed phenotype, fllB may perturb the complex network of low affinity cooperative interactions involved in the integrin-cytoskeleton association leading to an amplified defect in the transduction of mechanical forces across the plasma membrane. Alternatively, fllB may generate specific intracellular signals leading to reduced cell adhesion. These aspects deserve further investigation.

Dominant negative effects have been reported for arti- ficially mutated forms of adhesion receptors. A/31 integrin with a mutation in the extracellular domain that abolished ligand binding was shown to have a dominant negative effect on cell adhesion and spreading (Takada et al., 1992). More recently, a dominant negative effect has been described for a chimeric molecule containing the cytoplasmic domain of

the 131 integrin (Lukashev and Pytela, 1993). Expression of a truncated form of N-cadherin, a cell-cell adhesion receptor, consisting of the sole transmembrane and cytoplasmic regions, but lacking the extracellular domain resulted in reduced cell-cell adhesion (Kinter, 1992). The truncated form presumably competes with the native molecule by binding to cytoskeletal proteins necessary to transduce the mechanical force of adhesion. A second example comes from a chimeric membrane protein containing the cytoplasmic domain of desmoglein, a desmosomal cadherin (Troyanovsky et al., 1993). This chimera causes the disruption of desmosomes and detachment of intermediate filaments from the plasma membrane.

All examples discussed above consist of proteins with a functional intracellular segment and an altered or deleted extracellular domain. The data presented here show that an adhesion receptor containing an intact extracellular domain and an altered cytoplasmic region can also cause a dominant negative effect. In the case of/31B, in fact, correct extracellu- lax interactions elicit a peculiar intracellular response that leads to anomalous adhesion. The fact that/31B represents a splice variant raises the interesting question that cells may regulate their adhesive behavior by a splicing mechanism. The authors are indebted to F. Giancotti, R. Juliano, F. Sanchez-Madrid, and L. Zardi for the generous gift o f ill integrin antibodies. We also thank Rita Palumbo and Daniela Tomatis for helping in some of the experiments. The work was supported by grants from the Italian Association for Re- search on Cancer; from the Progetto Finalizzato "Applicazioni Cliniche della Ricerca Oncologica" of the National Research Council, and from the Biomed 1 of the European Community.

Received for publication 22 December 1993 and in revised form 28 June 1994.

References

Aibini, A., G. Allavena, A. Melchiorri, F. Giancotti, H. Richter, P. M. Comoglio, S. Parodi, G. R. Martin, and G. Tarone. 1987. Chemotaxis of 3T3 and SV3T3 cells to fibronectin is mediated through the cell attachment site in fibronectin and a fibronectin cell surface receptor. J. Cell Biol. 105: 1867-1872.

Altruda, F., P. Cervelia, G. Tarone, C. Botta, F. Balzac, G. Stefanuto, and L. Silengo. 1990. A human integrin ~1 with a unique cytoplasmic domain generated by alternative mRNA processing. Gene (Amst.). 95:261-266. Balzac, F., A. Belkin, V. Koteliansky, Y. Balabanow, F. Altruda, L Silengo,

and G. Tarone. 1993. Expression and functional analysis of a cytoplasmic domain variant of the ~1 integrin subunit. J. Cell Biol. 121:171-178. Burridge, C. A., K. Fath, T. Kelley, G. Nuckolls, and C. Turner. 1988. Focal

adhesions: transmembrane janctions between the extracelluiar matrix and the cytoskeleton. Annu. Rev. Cell Biol. 4:487-525.

Burridge, C.A., C. Turner, and L. Romer. 1992. Tyrosine phosphorylation of paxillin and pp125FAK accompanies cell adhesion to extracellular matrix: a role in cytoskeletal assembly. J. Cell Biol. 119:893-904.

Chamberlain, J. P. 1979. Fluorographic detection of radioactivity in polyacryl- amide gels with water-soluble fluor, sodium salicilate. Anal. Biochem. 98:!32-135.

Collo, G., L. Starr, and V. Quaranta. 1993. A new isoform of the laminin receptor integrin ~7/~1 is developmentally regulated in skeletal muscle. J.

Biol. Chem. 268:19019-19023.

Defilippi, P., L. Silengo, and G. Tarone. 1992. a t / ~ l integrin (iamin-receptor) is down-regulated by tumor necrosis factor B1 and intefleukin-1B in human endothelial cells. J. Biol. Chem. 267:18303-18307.

Defilippi, P., C. Bozzo, G. Volpe, G. Romano, M. Venturino, L. Silengo, and G. Tarone. 1994. Integrin-mediated signal transduction in human endothelial cells: analysis of tyrosine phosphorylation events. Cell Adhesion Commun. 2:75-86.

Ellis, L., E. Clauser, D. O. Morgan, M. Edery, R. A. Roth, and W. J. Rutter. 1986. Replacement of insulin receptor tyrosine residues 1162 and 1163 com- promise insulin-stimulated kinase activity and uptake of 2-deoxyglucose.

Cell. 45:721-732.

Engvall, E., and E. Ruoslahti. 1977. Binding of soluble form of fibroblast sur- face protein fibronectin to collagen. Int. J. Cancer. 20:1-5.

on October 12, 2005

www.jcb.org

(10)

C-eiger, B., and D. Ginsberg. 1991. The cytoplasmic domain of adherens-type junction. Cell Motil. Cytoskeleton. 20:1-6.

Giancotti, F. G., and E. Ruoslahti. 1990. Elevated levels of the ~5//~1 fibronectin receptor suppress the transformed phenotype of chinese hamster ovary cells. Cell. 60:849-859.

Guan, J. L., and D. Shalloway. 1992. Regulation of focal adhesion associated protein tyrosine kinase by both cellular adhesion and oncogenic transforma- tion. Nature (Lond.). 358:690-692.

Guan, J. L., J. E. Trevethick, and R. O. Hynes. 1991. Fibronectin/integrin in- teraction induces tyrosine phosphorylatlon of a 120 kD protein. Cell Regula-

tion. 2:951-964.

Hogervorst, F., I. Kuikman, A. G. Van Kessel, and A. Sonnenberg. 1991. Mo- lecular cloning of the human ~6 integrin subunit: alternative splicing of c~6 mRNA and chromosomal localization of the ~6 and/~4 genes. Fur. J. Bio-

chem. 199:425-433.

Horwitz, A., E. Duggan, C. Buck, M. C. Beckerle, and K. Burridge. 1986. Interaction of plasma membrane fibruuectin receptor with talin: a transmem- brane linkage. Nature (Lond.). 320:531-533.

Hynes, R. O. 1992. Integrins: versatility, modulation and signalling in cell adhesion. Cell. 69:11-25.

Jaken, S., K. Leach, and T. Klauck. 1989. Association of type3 protein kinaseC with focal contacts in rat embryo fibroblasts. J. Cell Biol. 109:697-704. Kinter, C. 1992. Regulation of embryonic cell adhesion by the cadherin cyto-

plasmic domain. Cell. 69:225-236.

Kornberg, L., H. S. Earp, J. T. Parsons, M. Schaller, andR. L. Juliano. 1992. Cell adhesion of integrin clustering increases phosphorylation of a focal adhesion associated tyrosine kinase. J. Biol. Chem. 267:23439-23442. Languino, L. R., and E. Ruoslahti. 1992. An alternative form of the integrin

/~1 subunit with a variant cytoplasmic domain. J. Biol. Chem. 267:7116- 7120.

Lukashev, M. E., and R. Pytela. 1993. Suppression of cell adhesion by autono- mously expressed C-terminal domain of the/~1 integrin subunit. Mol. Biol.

Cell. 4(Suppl.):2352.

Luna, E. J., and A. L. Hitt. 1992. Cytoskelcton-plasma membrane interactions

Science (Wash. DC). 258:955-964.

Otey, C. A., F. M. Pavalko, and K. Burridge. 1990. An interaction between c~-actinin and the E1 integrin subuhit in vitro. J. Cell Biol. 111:721-729. Reszka, A. A., R. A. Hayashi, and A. F. Horwitz. 1992. Identification of amino

acid sequences in the integrin /~1 cytoplasmic domain implicated in cytoskeletal association. J. Cell Biol. 117:1321-1330.

Rohrschneider, L. R. 1980. Adhesion plaques of Rous Sarcoma virus- transformed cells contain the src gene product. Proc. Natl. Acad. Sci. USA.

77:3514-3518.

Rossino, P., P. Defilippi, L. Silengo, and G. Tarone. 1991. Upregulation of integrin c~1//31 in human neuroblastoma cells differentiated by retinoic acid: correlation with increased neurite outgrowth response to laminin. Cell Regu-

lation. 2:1021-1033.

Ruoslahti E. 1991. Integrins. J. Clin. Invest. 87:1-5.

Sastry, S. K., and A. F. Horwitz. 1993. Integrin cytoplasmic domain: media- tors of cytoskeletal linkages and extra- and intraceUular initiated transmem- brane signalling. Curr. Opin. Cell Biol. 5:819-831.

Schaller, M. D., C. A. Borgman, B. S. Cobb, R. R. Vines, A. B. Reynolds, and J. T. Parsons. 1992. pp125FAK, a structurally unique protein tyrosine kinase associated with focal adhesion. Proc. Natl. Acad. Sci. USA. 89: 5192-5196.

Southern, P. J., and P. W. Berg. 1982. Transformation of mammalian cells to antibiotic resistance with a bacterial gene under control of the SV40 early region promoter. J. Mot.. Appl. Genet. 1:327-341.

Takada, Y., J. Yliinne, D. Mandelman, W. Puzon, and M. H. Ginsberg. 1992. A point mutation of integrin/~1 subunit blocks binding of ce5//~l to fibronec- tin and invasin but not recruitment to adhesion plaques. J. Cell Biol.

119:913-921.

Tamura, R. N., H. M. Cooper, G. Collo, and V. Quaranta. 1991. Cell type specific integrin variants with alternative ~ chain cytoplasmic domains.

Proc. Natl. Acad. Sci. USA. 88:10186-10187.

Tamura, R. N., C. Rozzo, L. Start, J. Chambers, L. F. Reichardt, H. M. Cooper, and V. Quaranta. 1990. Epithelial integrin ot6//g4: complete pri- mary structure of c~6 and variant forms of ~4. J. Cell Biol. 111:1593-1604. Troyanovsky, S. M., L. G. Eshkind, R. B. Troyanovsky, R. E. Leube, and W. W. Franke. 1993. Contribution of cytoplasmic domains of desmosomal cadherins to desmosome assembly and intermediate filament anchorage.

Cell. 72:561-574.

Turner, C. E., and K. Burridge. 1991. Transmembrane molecular assemblies in cell-extracellular matrix interactions. Curt. Opin. Cell Biol. 3:849-853. van Kuppevelt, T. H., L. Languino, J. O. Gailit, S. Suzuki, and E. Ruoslahti. 1989. An alternative cytoplasmic domain of the integrin/~3 subunit. Proc.

Natl. Acad. Sci. USA. 86:5415-5418.

Wang, N., J. P. Butler, and D. E. Ingberg. 1993. Mechanotransduction across the cell surface and through the cytoskeleton. Science (Wash. DC). 260: 1124-1127.

Yatuhgo, T., M. Irumi, H. Kasihiwagi, and M. Hayashi. 1988. Novel purifica- tion of vitronectin from human plasma by heparin affinity chromatography.

Cell Struct. Funct. 13:281-292.

Balzac et al. Role of BIB lntegrin in Cell Adhesion 565

on October 12, 2005

www.jcb.org